In a previous QUERI project our research group determined that more than 30% of Veterans with PTSD seen in VA facilities received a benzodiazepine prescription. Benzodiazepines remained the third most common medication prescribed to Veterans with PTSD. Analyses that controlled for co-existing Axis I disorders where benzodiazepines might be indicated found evidence of benzodiazepine prescribing in subgroups of PTSD patients where these medications are clearly contraindicated. Some members of a panel convened at the conclusion of our study remained divided on the evidence of the risk/benefit ratio for the use of benzodiazepines in PTSD patients in general, so we decided to focus this proposal on high-risk subgroups of patients with PTSD for whom the evidence is most clear; i.e., on those who have a co-occurring diagnosis where benzodiazepines are clearly contraindicated or those who because of demographic characteristics are most vulnerable to poor outcomes from chronic benzodiazepine use. Specifically, we chose to target reducing benzodiazepine use in subgroups of Veterans with a diagnosis of PTSD: 1) those with co-occurring substance use disorder, 2) those with a history of TBI, 3) older Veterans at risk of cognitive decline, 4) women Veterans at risk of increased psychotropic prescribing and 5) Veterans on opioids for chronic pain and benzodiazepines.
We had the following objectives:
(1) Develop decision support tools and obtain formative evaluation input from providers and patients.
(2) Use an academic detailing approach to test the decision support tools with providers.
(3) Conduct interviews with providers who participated in the intervention in Aim 2 to get their perspectives on facilitators and barriers as to the use of the tools with patients and academic detailing.
(4) Collect limited facility and provider level quantitative outcomes that include baseline prescribing rates and changes in prescribing patterns on specific psychotropics after the AD intervention is initiated.
This study utilized a mixed methods approach. For Aim 1 we obtained feedback from providers and patients through individual as well as group qualitative interviews. We interviewed 7 Veterans and 4 clinicians in the early development of the brochures and then interviewed 15 Veterans and 11 clinicians in later development. For Aim 2 we consented a total of 15 prescribing clinicians for the academic detailing intervention. For Aim 3 we interviewed 13 prescribing clinicians on the use of the decision support tools; 2 clinicians did not respond to interview requests and 1 had moved to an administrative position that no longer entailed direct patient care. For Aim 4, we used quantitative methodology to determine aggregate data of change in prescribing of benzodiazepines.
We chose to do this work in the VA Northern California Health Care System because that system had a trained academic detailing team and review of data suggested that there was a need to address the use of benzodiazepines in their management of PTSD. After delays in getting local IRB approval, we initiated the intervention only to learn that the experienced detailing pharmacist was moving. We waited for a new pharmacist to be hired, trained in detailing, and again initiated the intervention. The pharmacist hired was inexperienced, unfamiliar with the VA, and uncomfortable in the detailing role of meeting with prescribing clinicians to review their caseload and recommend changes.
What we learned is that decision support tools alone can be helpful to clinicians to do the right thing. Clinicians talked about the fact that it was helpful to give their patients the brochures and that they even used them to justify making changes to their prescribing patterns. They said that the brochures were a nice scapegoat; they told their patient it was a VA initiative and they really had to make changes in their prescribing. So that is encouraging and given the changes that we saw in prescribing, we think that just the dissemination of the brochures helped create a change in prescribing culture in the facilities.
In our analyses, we saw a significant improvement from baseline to the mid-point when the brochures were disseminated with 11 prescribing clinicians showing a decrease in the percentage of their PTSD patients who were on benzodiazepines. Only 2 clinicians showed an increase. However, our analyses also pointed to the need for sustainment strategies and this is where academic detailing would be helpful to encourage continued decreases in prescribing of benzodiazepines. We observed in 6 of our prescribing clinicians a small return to increased percentages of patients on benzodiazepines post-study. The rates did not return to baseline but did show a small return to their old prescribing patterns. This tells us that the use of decision support tools by the prescribing clinicians can be helpful but isn't sufficient to maintain change.
We think that the body of work (our two RRPs) supported by the MH QUERI has been tremendously effective in improving the state of prescribing in Veterans with PTSD as well as highlighting problems with the use of benzodiazepines.
External Links for this Project
Grant Number: I21HX001412-01A1
- Alexander B, Lund BC, Bernardy NC, Christopher ML, Friedman MJ. Early discontinuation and suboptimal dosing of prazosin: a potential missed opportunity for veterans with posttraumatic stress disorder. The Journal of clinical psychiatry. 2015 May 1; 76(5):e639-44. [view]
- Bernardy NC, Friedman MJ. Psychopharmacological strategies in the management of posttraumatic stress disorder (PTSD): what have we learned? Current psychiatry reports. 2015 Apr 1; 17(4):564. [view]